This is a relatively simplistic solution, but you could manually identify
sub-regions of your tracts-of-interest in MNI space (i.e., draw masks). Use
fnirt to warp your FA map to standard space, invert the warp fields and
align these masks to your native data. You could then convolve each
sub-region mask with your actual tract and then compute mean DTI metrics
within the new hybrid masks.
You would probably need to be very broad when defining the sub-regions in
standard space, to account for individual differences in the tracts that
have been determined in native space with probtrack. To help with this, you
could warp each subject's probtrack-defined tracts to MNI space, average
them, and use them as guides when tracing out this "standard/template set"
of
sub-regions that can be applied to the tracts in native space.
--Ryan
On Jan 27 2010, Martin Kavec wrote:
>Hi Michael,
>
> yeah exactly. I was struggling to send the image to the mailing list, but
> it
> was too big. That is exactly what I want. I know DTIStudio of Susumu Mori
> can
>give this profile, but I don't know how he does that.
>
>Although with cortico spinal track it is rather easy to go through slice
> direction, I am interested in more general solution, because the track of
> my
> interest is rather curvy. If you would take this approach with SLF you
> would
> mix FAs from rather distinct anatomical locations. Therefore something
> more
> tricky should be used. Something as David just suggested - strightening
> the
>track, but how to do it practically?
>
>Thanks,
>
>Martin
>
>On Wednesday 27 January 2010 20:42:46 Michael Scheel wrote:
>> Hi Martin,
>>
>> I've seen the paper by Sage et al. Quantitative diffusion tensor
>> imaging in
>> amyotrophic lateral sclerosis: Revisited See screenshot attached - is
>> this
>> what you want? I'd be interested in how to do this with fsl tools as
>> well.
>>
>> cheers, michael
>
|